摘要
为了提高复杂工业产品外观轮廓的合理性,需要高精度设计。产品外观合理性检测需要对设计产品的弯曲率、倾斜角度等复杂属性形成关联性的阀值,才能形成是否合理的判断。传统的人工视觉方法应用到复杂工业设计产品外观轮廓合理性检测时,对一些复杂的扭曲度变化、角度的非正常偏移等特征很难形成完成的判断规则,会产生检测误差大等弊端。提出了采用关联视觉检测技术的复杂工业设计产品外观轮廓合理性检测的方法,根据小波包分析法对设计产品的外观图像轮廓尺度特征进行分析与提取,再运用像素分类判断矩阵对尺度变化的合理性进行判断,完成复杂工业设计产品外观轮廓合理性检测。仿真结果表明,利用计算机视觉检测复杂工业设计产品外观轮廓合理性,精准性高,鲁棒性强。
In order to improve the rationality of product appearance and profile in some complex industrial design, high precision design is needed. A detection method for rationality of product appearance and profile in complex industrial design was proposed based on correlation visual inspection technology. According to wavelet packet analysis method, the scale features of designed product appearance and profile image were analyzed and extracted. Then, pixel classification judgment matrix was used to judge the rationality of the scale changes and complete the rationality detection of product appearance and profile in complex industrial design. Experimental results show that the detecting precision of the rationality of product appearance and profile in complex industrial design is high, and the robustness is strong.
出处
《计算机仿真》
CSCD
北大核心
2015年第4期355-358,362,共5页
Computer Simulation
关键词
计算机视觉
小波包分析
产品外观合理性
Computer vision
Wavelet packet analysis
Rationality of product appearance
